What this calculator does
Relative risk reductions sound impressive and tell you very little. A treatment that cuts risk by 25 per cent might prevent one event in four people or one in four hundred, depending entirely on how common the event was to begin with.
Number needed to treat answers the question directly. Cutting an event rate from 20 to 15 per cent gives an NNT of 20.0, meaning twenty people must be treated for one to benefit. Cutting it from 20 to 19 gives an NNT of 100.
The formula
The absolute risk reduction is the difference between the two event rates. NNT is one divided by that difference, expressed as a proportion.
| Term | Meaning |
|---|---|
| Absolute risk reduction | The difference in event rates, in percentage points. |
| NNT | One divided by that difference: patients treated per additional benefit. |
| NNH | Number needed to harm, the same calculation when a treatment increases a bad outcome. |
The inputs explained
| Field | What to enter |
|---|---|
| Control group event rate (%) | The event rate in the control group, as a percentage. |
| Treatment group event rate (%) | The event rate in the treatment group, as a percentage. |
When to use it
Reading a trial result
NNT converts a relative reduction into something concrete about patients.
Comparing two treatments
A lower NNT means fewer people treated for each one who benefits.
Weighing benefit against harm
NNT against NNH puts benefit and side effects on the same scale.
Worked examples
Every figure in the tables below is produced by this page’s own calculator at build time, so the numbers and the tool always agree. Select any row to load that scenario.
How does the size of the reduction change NNT?
Four treatment event rates against the same control.
Questions
Why is NNT better than relative risk reduction?
Because relative reduction hides how common the outcome was. Halving a risk from 20 per cent to 10 helps one person in ten; halving it from 0.2 per cent to 0.1 helps one in a thousand. Both are "50 per cent" reductions.
What is a good NNT?
It depends entirely on the outcome and the treatment. An NNT of 100 is excellent for a cheap, safe drug preventing death, and poor for an expensive one with serious side effects preventing something minor.
Does NNT depend on the time period?
Very much. An NNT is only meaningful alongside the duration over which it was measured, since event rates accumulate. An NNT of 20 over five years is quite different from an NNT of 20 over five weeks.
Should this be rounded?
Up, to a whole person, since you cannot treat a fraction of a patient. An NNT of 20.4 is reported as 21 in practice.
For efficacy from attack rates, see the vaccine efficacy calculator. For a general health baseline, see the BMI calculator.